sign in sign up
<<
Home , Dimorphodon, Rhamphorhynchus, Scaphognathus

University of Birmingham

Dietary diversity and evolution of the earliest flying vertebrates revealed by dental microwear texture analysis Bestwick, Jordan; Unwin, David; Butler, Richard; Purnell, Mark

DOI: 10.1038/s41467-020-19022-2 License: Creative Commons: Attribution (CC BY)

Document Version Publisher's PDF, also known as Version of record Citation for published version (Harvard): Bestwick, J, Unwin, D, Butler, R & Purnell, M 2020, 'Dietary diversity and evolution of the earliest flying vertebrates revealed by dental microwear texture analysis', Nature Communications, vol. 2020, 5293. https://doi.org/10.1038/s41467-020-19022-2

Link to publication on Research at Birmingham portal

Publisher Rights Statement: Bestwick, J., Unwin, D.M., Butler, R.J. et al. Dietary diversity and evolution of the earliest flying vertebrates revealed by dental microwear texture analysis. Nat Commun 11, 5293 (2020). https://doi.org/10.1038/s41467-020-19022-2

General rights Unless a licence is specified above, all rights (including copyright and moral rights) in this document are retained by the authors and/or the copyright holders. The express permission of the copyright holder must be obtained for any use of this material other than for purposes permitted by law.

•Users may freely distribute the URL that is used to identify this publication. •Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private study or non-commercial research. •User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and Patents Act 1988 (?) •Users may not further distribute the material nor use it for the purposes of commercial gain.

Where a licence is displayed above, please note the terms and conditions of the licence govern your use of this document.

When citing, please reference the published version. Take down policy While the University of Birmingham exercises care and attention in making items available there are rare occasions when an item has been uploaded in error or has been deemed to be commercially or otherwise sensitive.

If you believe that this is the case for this document, please contact [email protected] providing details and we will remove access to the work immediately and investigate.

Download date: 01. Oct. 2021 ARTICLE

https://doi.org/10.1038/s41467-020-19022-2 OPEN Dietary diversity and evolution of the earliest flying vertebrates revealed by dental microwear texture analysis ✉ ✉ Jordan Bestwick 1,2 , David M. Unwin3, Richard J. Butler 2 & Mark A. Purnell 1

Pterosaurs, the first vertebrates to evolve active flight, lived between 210 and 66 million ago. They were important components of Mesozoic ecosystems, and reconstructing pter- 1234567890():,; osaur diets is vital for understanding their origins, their roles within Mesozoic food webs and the impact of other flying vertebrates (i.e. ) on their evolution. However, dietary hypotheses are poorly constrained as most rely on morphological-functional analo- gies. Here we constrain the diets of 17 pterosaur genera by applying dental microwear texture analysis to the three-dimensional sub-micrometre scale tooth textures that formed during food consumption. We reveal broad patterns of dietary diversity (e.g. as a vertebrate consumer; Austriadactylus as a consumer of ‘hard’ invertebrates) and direct evi- dence of sympatric niche partitioning ( as a piscivore; Pterodactylus as a generalist invertebrate consumer). We propose that the ancestral pterosaur diet was dominated by invertebrates and later evolved into piscivores and carnivores, shifts that might reflect ecological displacements due to pterosaur- competition.

1 Centre for Palaeobiology Research, School of Geography, Geology and the Environment, University of Leicester, Leicester LE1 7RH, UK. 2 School of Geography, Earth and Environmental Sciences, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK. 3 Centre for Palaeobiology Research, School of ✉ Museum Studies, University of Leicester, Leicester LE1 7RF, UK. email: [email protected]; [email protected]

NATURE COMMUNICATIONS | (2020) 11:5293 | https://doi.org/10.1038/s41467-020-19022-2 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19022-2

terosaurs, the first group of vertebrates to evolve active a range of different prey types. We also provide evidence of flight, lived between 210 and 66 million years ago during sympatric niche partitioning between Rhamphorhynchus and P 1 2 the Mesozoic . They had a global distribution and inhab- Pterodactylus, and evidence of ontogenetic niche partitioning in ited a range of terrestrial, coastal and marine environments3.As Rhamphorhynchus. Furthermore, we reconstruct pterosaur diet- important components of Mesozoic biotas4, robust reconstruc- ary evolution from our DMTA results by projecting pruned, time- tions of pterosaur diets are thus crucial not only for under- calibrated trees from three pterosaur phylogenies into the extant standing the ecological roles they performed within Mesozoic multivariate framework, and by mapping pterosaur food webs, but for helping address broader ecological and evo- microwear characteristics from the reptile multivariate frame- lutionary debates that enhance our understanding of Mesozoic work onto each tree. We find that the diets of the first pterosaurs ecosystems. Key debates include: whether sympatric of largely consisted of invertebrates and that pterosaurs became pterosaur competed for, or partitioned, food resources5; whether increasingly carnivorous and piscivorous over evolutionary time. ontogenetic stages of a single species competed for, or partitioned, These results provide insight into the multitude of roles that resources and thus had life-histories more similar to those of pterosaurs performed within Mesozoic food webs and highlight extant birds or non-avian reptiles6–8; and what impact the the applicability of DMTA for investigating extinct ecosystems. emergence and radiation of birds had on pterosaur dietary evo- 9 lution , and vice versa. Results and discussion Pterosaur diets have been the subject of considerable investi- Extant reptile and microwear frameworks. The extant rep- gation, but for many taxa there is little consensus regarding their tile framework comprises six crocodilian and seven monitor diet4. Dietary studies are inherently difficult because pterosaurs lizard species assigned to one of five dietary guilds: carnivores have no modern descendants10, and many dietary hypotheses are (tetrapod consumers); ‘harder’ invertebrate consumers (e.g. bee- based on weakly supported analogies between the morphology tles, crustaceans and shelled gastropods); ‘softer’ invertebrate and function of anatomical structures in extant and consumers (e.g. grasshoppers); omnivores; and piscivores (fish pterosaurs4. Fossilised stomach contents can be informative but consumers)18 (Supplementary Fig. 1 and Supplementary Data 1 are limited to a small number of specimens from a few species4,11. and 2). See Fig. 1a–f for examples of digital elevation models of Furthermore, studies that include dietary hypotheses typically extant reptile and pterosaur tooth surfaces from which texture focus on just one or two taxa, and the absence of a multi-taxon data were acquired. As previously reported18, four texture para- framework prevents reliable investigations into resource parti- meters differ significantly between reptile guilds (ANOVA; Sup- tioning between sympatric pterosaurs and the role of diet in plementary Data 3 and 4) and principal components analysis evolutionary transitions and responses to environmental separates them in a texture-dietary space defined by PC axes 1 upheavals12,13. and 2 (Fig. 2 and Supplementary Fig. 2). PC 1 negatively corre- A more robust approach to understanding diet involves dental lates with proportions of total vertebrates in reptile diets (r = microwear texture analysis (DMTA)—quantitative analysis of the s −0.3564, P = 0.0004) but positively correlates with total inver- sub-micrometre scale three-dimensional textures that form on – tebrates (r = 0.3192, P = 0.0016), while PC 2 positively correlates tooth surfaces during food consumption14 17. The difficulty s with dietary proportions of ‘softer’ invertebrates (r = 0.2907, P = experienced by consumers in piercing and chewing food items s 0.0043; Fig. 2; Supplementary Fig. 3 and Supplementary Data 5). determines the microwear patterns that form on tooth crowns In very broad terms, ‘harder’ foods correlate with rougher tex- which, consequently, provides direct evidence of diet18. Analyses tures, see ref. 18 for details). use standardised texture parameters14,19 to quantify microwear, The extant bat framework comprises eight species assigned to and thus dietary, differences between species and/or populations one of four guilds: carnivores; ‘harder’ invertebrate consumers; and therefore do not assume direct relationships between the piscivores; and ‘softest’ invertebrate (e.g. moths) consumers morphology and inferred functions of teeth14,15,17. Recent work (Supplementary Fig. 1, Supplementary Data 1 and 2 and has demonstrated that DMTA of non-occlusal (non-chewing) Supplementary Note 1). PCA of the 15 texture parameters that tooth surfaces of extant , including archosaurs (the significantly differ between guilds (Supplementary Data 3 and 4) to which pterosaurs belong), differs between dietary guilds18,20. separates guilds in a texture-dietary space defined by PC axes 1 This relationship between microwear texture and diet in taxa that and 2 (Supplementary Fig. 2). PC 1 positively correlates with do not chew food items provides a robust multivariate framework proportions of total invertebrates, ‘softest’ invertebrates, plant for this study, allowing us to quantitatively test and constrain matter and dietary generalism, and negatively correlates with total pterosaur dietary hypotheses, and to explore dietary shifts within vertebrates and tetrapods. PC 2 positively correlates with dietary this clade across evolutionary time. proportions of tetrapods (Supplementary Fig. 3, Supplementary Here, we apply DMTA to the non-occlusal tooth surfaces of 17 Data 5 and Supplementary Note 1). pterosaur genera. These pterosaurs span the first 120 million years of pterosaur evolution from the Upper (Norian) to the Upper (Cenomanian). We use DMTA to test three Pterosaur dietary reconstructions. Projecting pterosaur data into hypotheses: that pterosaur teeth preserve evidence of dietary the reptile texture-dietary space plots them within the bounds of differences between taxa, between sympatric species and between the reptiles, and where pterosaur taxa are represented by multiple ontogenetic stages of the same species. Our hypothesis testing specimens, there is tendency for them to cluster together rather uses the multivariate framework provided by DMTA of extant than exhibit a random distribution across the entire texture- reptiles18 as well as additional, independent tests using a DMTA dietary space (e.g. Darwinopterus, Dimorphodon) although some framework derived from the non-occlusal tooth surfaces of , are more broadly distributed (e.g. , Pterodactylus; the only extant group of dentulous flying vertebrates (given the Fig. 2; Supplementary Fig. 2 and Supplementary Note 2). This differences between bats, extant reptiles and pterosaurs, con- non-random distribution is noteworthy, given that pterosaur data silience in the results of these independent analyses represents a played no role in structuring the PCA. The independently derived very stringent test of our approach). Our results allow us to texture-dietary space for bats gives similar results (Supplementary propose a number of specific dietary interpretations, including Fig. 3 and Supplementary Note 2). In both analyses, PC 1 explains carnivores, piscivores and invertebrates consumers, and to dis- the majority of microwear variation between reptiles and bats tinguish between dietary specialists and pterosaurs that consumed (66.2% and 55% respectively) and is therefore the more

2 NATURE COMMUNICATIONS | (2020) 11:5293 | https://doi.org/10.1038/s41467-020-19022-2 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19022-2 ARTICLE

a b c

Height (µm) 3 2.25 Piscivore (Gavialis gangeticus)‘Harder’ invertebrate consumer Omnivore (Varanus olivaceus) (Crocodylus acutus) 1.5 d ef0.75 0 –0.75 –1.5 –2.25 –3 –3.75

Istiodactylus latidens robustus Austriadactylus cristatus

Fig. 1 Example scale-limited tooth surfaces of reptile dietary guilds and pterosaurs. a–c Reptile dietary guilds; a piscivore (Gavialis gangeticus; gharial), b ‘harder’ invertebrate consumer (Crocodylus acutus; American crocodile) and c omnivore (Varanus olivaceus; Grey’s monitor lizard). d–f Pterosaurs; d (PCA plot number 16 in Supplementary Fig. 2), e Coloborhynchus (PCA number 5) and f Austriadactylus (PCA number 2). Measured areas 146 × 110 µm in size. Topographic scale in micrometres. Skull diagrams of extant reptiles and pterosaurs not to scale (see ‘Methods' for sources). informative axis for interpreting and constraining pterosaur diets piscivore or a carnivore4 based on its razor-edged, lancet-shaped, (this does not negate the usefulness of PC 2 as this axis can interlocking teeth, interpreted as an adaptation for defleshing provide additional constraints, see below). PC 1 approximates a carcasses25. Our DMTA results place Istiodactylus at the extreme spectrum from invertebrate dominated diets at one end to ver- end of the vertebrate scale with respect to reptile-based PC 1, tebrate dominated diets at the other. Pterosaur taxa with values plotting closest to a number of reptile carnivore samples (Fig. 2). similar to vertebrate consuming guilds in the reptile-based ana- This strong concordance between multiple independent lines of lysis also plot with vertebrate consuming guilds in the bat-based evidence gives us confidence in our DMTA results and provides analysis. The same is true for those with PC 1 values similar to further, quantitative evidence that Rhamphorhynchus was a invertebrate consuming guilds (Fig. 2; Supplementary Figs. 2 and piscivore and that Istiodactylus was an obligate vertebrate 3 and Supplementary Note 2). That these independent analyses consumer, most likely a carnivore. produce similar results provides powerful evidence that the dif- In contrast to the few pterosaur taxa where dietary ferences between pterosaur non-occlusal tooth textures are hypotheses are relatively uncontroversial, DMTA gives us attributable to dietary differences. This allows us to test previous insight into pterosaurs that have poorly constrained diets. dietary hypotheses, provide more refined characterisation of the Dimorphodon, for example, has been variably interpreted as ecological roles that pterosaurs performed within food webs, and carnivorous, piscivorous and insectivorous based on compara- explore their dietary evolution. For simplicity we focus here on tive anatomy of the skull with extant birds and crocodilians (see the reptile-based analysis (see Supplementary Note 1 for details of ref. 4 and references therein). The high degree of overlap of the bat-based analysis). Dimorphodon tooth textures with reptile carnivores and Further support for the robustness of our approach comes piscivores across PCs 1 and 2 (Fig. 2) suggests that the diet of from consilience with other, well-supported dietary interpreta- Dimorphodon largely consisted of vertebrates, although some tions. While decades of research have yielded little consensus consumption of ‘softer’ invertebrates cannot be entirely ruled regarding the diet of most pterosaurs, there are a few taxa for out. Lonchodraco and Serradraco are poorly known pterosaurs which multiple lines of evidence provide more secure hypotheses. whose anatomy and ecology is largely speculative1,4. Serradraco For example, multiple specimens of Rhamphorhynchus are is located towards the extreme end of the reptile carnivore/ preserved in association with fish remains as gut contents, and piscivore scale and overlaps more strongly with reptile their sharp conical teeth have been compared to extant piscivores along PC 2 (Fig. 2). This suggests that Serradraco piscivores21–24, a dietary hypothesis with which most analyses was predominantly piscivorous. Lonchodraco exhibits a very agree4. In our analysis, all but one of the 14 specimens overlap similar PC 2 value to Serradraco and a more positive value along with reptile piscivores along PCs 1 and 2 (Fig. 2), and the PC 1 (Fig. 2) which together suggests the inclusion of some distribution of Rhamphorhynchus corresponds more closely to invertebrates within a largely piscivorous diet. the convex hull for piscivores than any other group (Fig. 2). DMTA results challenge some previous dietary hypotheses. Similarly, Istiodactylus has consistently been reconstructed as a Austriadactylus, for example, has been interpreted as carnivorous

NATURE COMMUNICATIONS | (2020) 11:5293 | https://doi.org/10.1038/s41467-020-19022-2 | www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19022-2

3 ‘Softer’ invertebrate consumers

2 Carnivores 1

c 0 Omnivores a –1

Piscivores –2 Principal component 2 (25.1%)

b –3 ‘Harder’ invertebrate consumers Increasing proportion of total vertebrates in diet

Increasing proportion of total invertebrates in diet Increasing proportion of ‘softer’ invertebrates in diet –4 –4–3–2–10123456 Principal component 1 (66.2%)

Anhanguera Darwinopterus Haopterus Pterodactylus Austriadactylus Dimorphodon Istiodactylus Rhamphorhynchus Boreopterus Dorygnathus Jiangchangnathus Campylognathoides Lonchodraco Serradraco Coloborhynchus

Fig. 2 Quantitative textural analysis of microwear in extant reptiles and pterosaurs. Texture-dietary space of International Organisation for Standardisation (ISO) texture parameters for reptiles and pterosaurs. Texture-dietary space based on extant reptile data (n = 95) with pterosaurs (n = 40) projected onto the first two axes as unknown datum points. Extant specimens with associated letters represent surfaces a–c in Fig. 1. Arrows show significant correlations of dietary characteristics along PC axes 1 and 2. Pterosaur specimen information corresponding to PCA plot number can be found in Supplementary Fig. 2 and in Supplementary Data 1. Texture-dietary space adapted from Fig. 2 of ref. 18 (https://www.nature.com/articles/s41598-019- 48154-9/figures/2) by Jordan Bestwick and Mark Purnell under a Creative Commons Attribution 4.0 International License https://creativecommons.org/ licenses/by/4.0/ to include the pterosaur data. based on functional analyses of the closing mechanics of its had relatively strong bite forces which would have facilitated jaws26. However, its position in the texture-dietary space, with dietary generalism29. It is difficult to determine unequivocal high values for PC 1, suggests a diet dominated by invertebrates dietary preferences for Solnhofen species represented by single (Fig. 2). Coloborhynchus has been interpreted as piscivorous specimens that fall in the centre of the texture-dietary space, but based on its tooth morphology24 and on theoretical modelling of Scaphognathus and Germanodactylus may have consumed both possible fishing behaviours27; in the texture-dietary space its PC 1 vertebrates and invertebrates (Fig. 2), with evidence from the value is comparable to carnivores and piscivores but its positive analysis of bat diets suggestive of more invertebrates (Supple- PC 2 value suggests it also consumed a high proportion of ‘softer’ mentary Fig. 3). While this might indicate finer niche partitioning invertebrates (Fig. 2). It is unlikely that Coloborhynchus was a between Solnhofen pterosaurs, it is also likely that these species specialist piscivore; our analysis points towards a broader dietary were not typical members of the community as they are relatively range than has previously been proposed. rare (<5 specimens in both cases). Ontogenetic dietary shifts have been suggested for several pterosaurs6–8,30,31 but the hypothesis lacks evidential support. In Sympatric partitioning and ontogenetic dietary shifts. Our our study, the dietary separation of Rhamphorhynchus specimens results also have a bearing on contentious hypotheses of ecolo- along PC 1 correlates with specimen size (reptile PCA; lower jaw = − = gical interaction and competition between taxa: did sympatric length; rs 0.6044, P 0.0221; Fig. 3 and Supplementary 5,24,28 = − pterosaurs compete for or partition food resources ? DMTA Note 2). PC 2 values do not correlate with size (rs 0.3451, demonstrates niche partitioning between pterosaurs from the P = 0.2269). Specimens that most likely represent highly Solnhofen Limestones. Rhamphorhynchus and Pterodactylus, taxa immature individuals based on their size and degree of skeletal with the largest sample sizes, exhibit separation along PC 1 ossification6 exhibit positive PC 1 values, indicating invertebrate- (reptile PCA: t = −2.431, d.f. = 18, P = 0.0257; Fig. 2; see Sup- dominated diets. Specimens that represent late-stage juveniles plementary Note 2 for similar results in the bat-based analysis), and adults exhibit decreasing PC 1 values and thus increasingly but not along PC 2 (t = −1.303, d.f. = 18, P = 0.2087). More piscivorous diets (Fig. 2). Ontogenetic dietary shifts are common negative PC 1 values for Rhamphorhynchus indicate piscivory, in extant reptiles where offspring usually feed themselves (e.g. while Pterodactylus individuals exhibit a broader range of more ref. 32), but are not exhibited by birds8. DMTA thus provides positive PC 1 values, indicating invertebrate-dominated, and direct evidence of ontogenetic dietary shifts in a pterosaur, and possibly more generalist diets (Fig. 2). This agrees with theoretical supports the hypothesis that pterosaur life-histories were more – models of pterosaur feeding behaviours that suggest Pterodactylus like those of reptiles than birds or bats6 8,30.

4 NATURE COMMUNICATIONS | (2020) 11:5293 | https://doi.org/10.1038/s41467-020-19022-2 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19022-2 ARTICLE

3 2 R = 0.511 a 3 Time from root (ma) 0 107.8 ‘Softer’ invertebrate consumers

2 5 2 Carnivores

1 22 16 11 39 2 34 8 1 0 3 4 omnivores 14 40 18 –1 1 7 0 15 17 Piscivores

Principal component 2 –2 Principal component 1

–1 –3 ‘Harder’ invertebrate consumers Increasing proportion of total vertebrates in diet Increasing proportion of total vertebrates in diet

Increasing proportion of total invertebrates in diet Increasing proportion of total invertebrates in diet –4 Increasing proportion of ‘softer’ invertebrates in diet –4 –3 –2 –1 0 1 2 3 4 5 6 –2 Principal component 1 0 25 50 75 100 125 150 175 Lower jaw length (mm) b Campylognathoides Fig. 3 Rhamphorhynchus ontogenetic shifts in the reptile texture-dietary Lonchodraco space. Regression of lower jaw length (n = 14), as a proxy for specimen size, plotted against PC 1 values from the extant reptile texture-dietary Serradraco space. The solid grey line denotes the line of best fit and grey shading Germanodactylus denotes the 95% confidence intervals. Skull diagrams represent scaled Boreopterus examples of an adult, juvenile and hatchling Rhamphorhynchus, with the grey dashed lines denoting their relative size. These diagrams illustrate the size Haopterus disparity between sampled life-history stages and are not specimen Istiodactylus reconstructions of the datum points they are associated with. See Coloborhynchus Supplementary Fig. 4 for Rhamphorhynchus ontogenetic shifts in the bat texture-dietary space and the ‘Methods' for the skull diagram sources. Pterodactylus

Pterosaur dietary evolution. The implications of our results for Darwinopterus pterosaur dietary evolution were explored in two ways. First, by Rhamphorhynchus independently projecting pruned, time-calibrated trees from three pterosaur phylogenies, Lü et al.33, Andres and Myers34, and Dorygnathus Wang et al.35 into the first two axes of the reptile texture-dietary Jiangchangnathus ‘ space, creating phylo-texture-dietary spaces' analogous to phy- Scaphognathus lomorphospaces. Second, by mapping pterosaur PC scores from the first two reptile texture-dietary axes onto each tree. The Austriadactylus PC 1 value –3.37 2.58 results using the Lü et al.33 phylogeny (Fig. 4, Supplementary Dimorphodon Figs. 5, 6 and 9 and Supplementary Data 6) are presented here as this phylogeny exhibits higher stratigraphic congruence than Fig. 4 Pterosaur dietary evolution. a Phylo-texture-dietary space of alternatives36 (see Supplementary Figs. 7 and 10, Supplementary pterosaur microwear from projecting a time-calibrated, pruned tree from Lü Data 6 and Supplementary Note 3 for results using the Andres et al.33 onto the first two PC axes of the extant reptile texture-dietary and Myers34 phylogeny, and Supplementary Figs. 8 and 11, space. b Ancestral character-state reconstruction of pterosaur dietary Supplementary Data 6 and Supplementary Note 3 for results evolution from mapping pterosaur PC 1 values onto a time-calibrated, using the Wang et al.35 phylogeny). PC 1 values of the pterosaur pruned tree from Lü et al.33. To account for ontogenetic changes in diet, specimens included for dietary evolution reconstructions nega- only the largest specimen of respective pterosaur taxa, identified by lower = − = jaw length, were included. Pterosaur symbols same as Fig. 2. Skull diagrams tively correlate with stratigraphic age (rs 0.4859, P 0.048) = = of well-preserved pterosaurs not to scale (see ‘Methods' for sources). and PC 2 values do not correlate with age (rs 0.0552, P 0.8333). There is a significant phylogenetic signal in pterosaur Phylo-texture-dietary space adapted from Fig. 2 of ref. 18 (https://www. microwear along PC 1 (K = 0.9368, P = 0.031, λ = 0.9999, P = nature.com/articles/s41598-019-48154-9/figures/2) by Jordan Bestwick 0.000143), but no signal along PC 2 (K = 0.262, P = 0.775, λ = under a Creative Commons Attribution 4.0 International License https:// 0.201, P = 0.4014). The ancestral PC 1 value estimate of Pter- creativecommons.org/licenses/by/4.0/ to remove the extant reptile osauria (node 1, Supplementary Fig. 9) is 2.2 (Fig. 4b), and the datum points and to include the pterosaur data and phylogeny. ancestral PC 2 estimate is 0.428 (Supplementary Fig. 5 and Supplementary Data 6). This suggests that the ancestral pterosaur consumers (Fig. 4b). Interestingly, the invertebrate-dominated diet was invertebrate-dominated. Over time, pterosaurs move diet of Pterodactylus and the possible inclusion of invertebrates in through phylo-texture-dietary space to occupy increasingly the diet of Lonchodraco are both secondarily derived (Fig. 4b). negative PC 1 values, reflecting a shift towards consumption of These results provide quantitative evidence that pterosaurs more vertebrates. This is most pronounced in members of initially evolved as invertebrate consumers before expanding into Istiodactylidae and Lonchodectidae (nodes 13 and 16 respectively, piscivorous and carnivorous niches12,13. The causes of this shift Supplementary Fig. 9), which have estimated ancestral PC 1 towards vertebrate-dominated diets require further investigation, values of −2.137 and −2.007, respectively (Supplementary but might reflect ecological interactions with other taxa that Data 6) and independently evolved into obligate vertebrate radiated through the Mesozoic12,13. Specifically, competition with

NATURE COMMUNICATIONS | (2020) 11:5293 | https://doi.org/10.1038/s41467-020-19022-2 | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19022-2 birds, which first appeared in the Upper and diversified proposed for dentulous pterosaurs4. In reptiles, Crocodylus porosus adults in the Lower Cretaceous, has been invoked to explain the (saltwater crocodile, n = 6), Varanus komodoensis (Komodo dragon, n = 4), Varanus nebulosus (clouded monitor, n = 11), Varanus rudicollis (roughneck decline of small-bodied pterosaurs, but this hypothesis is = = 5,9,30,37 monitor, n 8), and Varanus salvator (Asian water monitor, n 8) were assigned controversial . DMTA provides an opportunity for testing to the carnivore guild (total n = 37); Crocodylus acutus (American crocodile, n = 7) hypotheses of competitive interaction upon which resolution of and Crocodylus porosus juveniles (n = 5) were assigned to the ‘harder’ invertebrate this ongoing debate will depend. consumer guild (total n = 12); Varanus niloticus (Nile monitor, n = 8) and = ‘ ’ In summary, our analyses provide quantitative evidence of Varanus prasinus (emerald tree monitor, n 7) were assigned to the softer invertebrate consumer guild (total n = 15); Varanus olivaceus (Gray’s monitor, n pterosaur diets, revealing that dietary preferences ranged across = 6) was assigned to the omnivore guild (total n = 6); and mississippiensis consumption of invertebrates, carnivory and piscivory. This has (American alligator, n = 8), Caiman crocodilus (spectacled caiman, n = 6), allowed us to explicitly constrain diets for some pterosaurs, Crocodylus niloticus (Nile crocodile, n = 4), and Gavialis gangeticus (gharial, n = 7) enabling more precise characterisations of pterosaurs’ roles were assigned to the piscivore guild (total n = 25). In bats, Trachops cirrhosis (fringe-lipped bat, n = 8) and Vampyrum spectrum (spectral bat, n = 9) were within Mesozoic food webs and providing insight into pterosaur assigned to the carnivore guild (total n = 17); Nyctalus noctula (noctule bat, n = 7) niche partitioning and life-histories. Our study sets a benchmark and Rhinolophus ferrumequinum (greater horseshoe bat, n = 8) were assigned to for robust interpretation of extinct reptile diets through DMTA of the ‘harder’ invertebrate consumer guild (total n = 15); Myotis capaccini (long- non-occlusal tooth surfaces and highlights the potential of the fingered bat, n = 7), Myotis mystacinus (whiskered bat, n = 6), and Plecotus auritus (brown long-eared bat, n = 6) were assigned to the ‘softest’ invertebrate consumer approach to enhance our understanding of ancient ecosystems. guild (total n = 19); and Noctilio leporinus (greater bulldog bat, n = 8) was assigned to the piscivore guild (total n = 8). See Fig. S1 for an overview of how reptiles and Methods bats were assigned to guilds. The Cr. acutus skull diagram in Fig. 1 was drawn from UF 54201, the G. gangeticus diagram from Wikimedia Commons under the Specimen material. Tooth microwear textures were sampled from 17 pterosaur Creative Commons Attribution-Share Alike 2.0 Generic license (https:// genera. Eight bat species were sampled and the microwear texture data of 13 reptile creativecommons.org/licenses/by-sa/2.0/deed.en), and the V. olivaceus diagram species, comprising six crocodilians and seven monitor lizards, from ref. 18 were from UF 57207. included to serve as extant multivariate frameworks. Crocodilians and monitor lizards were chosen to provide the primary framework because their dentitions have been compared with those of pterosaurs12, and they often forage on the Sampling strategy. Pterosaur specimens were cleaned before sampling using an margins of terrestrial and aquatic environments38,39, comparable to hypothesised ethaline solvent gel, produced, and applied according to Williams and Doyle63. pterosaur foraging behaviours40. Bats were chosen as an independent comparator Extant reptile and bat teeth from dry skeletal specimens were cleaned using 70% framework because they are the only extant group of dentulous flying vertebrates, ethanol-soaked cotton swabs to remove dirt and consolidant. Microwear data were and their foraging behaviours have sometimes been extrapolated to pterosaurs (see acquired from non-occlusal (non-chewing) labial surfaces, as close to the tooth ref. 1 and references therein). Extant and specimens were sampled from the apex as possible. In extant reptiles, the mesial-most dentary tooth was sampled; in Bayerische Staatsammlung für Paläontologie und Geologie, Munich, Germany pterosaurs, the mesial-most tooth of the premaxilla or dentary tooth, based on (BSPG); Field Museum of Natural History, Chicago, USA (FMNH); British Geo- preservation quality, was sampled. No preference was given to the left or right logical Survey, British Geology Collections, Keyworth, UK (GSM); Institute of tooth in extant reptiles and teeth were pooled in analyses. Wear facets that likely Vertebrate and Paleoanthropology, Beijing, China (IVPP); Grant formed from tooth–tooth occlusion from the opening and closing of jaws, char- Museum of Zoology, University College London, UK (LDUCZ); Museum für acterised by their vertical orientation, elliptical shape and parallel features12,64, Naturkunde der Humboldt-Universität Berlin, Berlin, Germany (MB); Natural were not sampled. Bat microwear data were acquired from the non-occlusal labial History Museum, London, UK (NHMUK); University of Oxford Museum of surface of the canine in the dentary, as close to the apex as possible. Canines were Natural History, Oxford, UK (OUMNH); Paleontological Museum of Liaoning, sampled over premolars and molars because they represent closer functional Shenyang, China (PMOL); Staatliches Museum für Naturkunde, Karlsruhe, Ger- analogues to reptile teeth since they are not used for chewing. No preference was many (SMNK); Staatliches Museum für Naturkunde, Stuttgart, Germany (SMNS); given to the left or right canine. To test for ontogenetic dietary shifts in Rham- Teylers Museum, Haarlem, Netherlands (TM); Royal Tyrrell Museum of phorhynchus, lower jaws were measured in lateral view from the anterior tip of the Palaeontology, Drumheller, Canada (TMP); Florida Museum of Natural History, dentary to the posterior margin of the surangular. Lower jaw length is a reliable Gainesville, USA (UF); and the National Museum of Natural History, Smithsonian proxy for overall specimen size as Rhamphorhynchus skulls exhibit strongly linear Institute, Washington DC, USA (USNM). See Supplementary Data 1 for the size relationships with respect to other anatomical structures, such as the humerus complete specimen list. and wing finger6, and we follow this analysis in our use of ‘adult’, ‘juvenile’ and ‘hatchling’. The example adult and hatchling Rhamphorhynchus skull diagrams in Fig. 3 were redrawn from Bennett6, and the example juvenile diagram was traced Dietary guild assignments. Reptiles and bats were selected to include taxa with from ref. 4 under a Creative Commons Attribution 4.0 International License well-constrained dietary differences determined from stomach and/or faecal con- fi 38,41–60 (https://creativecommons.org/licenses/by/4.0/). High delity moulds were taken of tent studies (Supplementary Data 2). Studies were chosen that met as many teeth using President Jet Regular Body polyvinylsiloxane (Coltène/Whaledent Ltd., of the following criteria as possible: representative sample sizes, dietary composi- Burgess Hill, West Sussex UK). Initial moulds taken from each specimen were tions represented as volumetric data (or frequency data at an absolute minimum); discarded to remove any remaining dirt and all analyses were performed on second and spatial proximity of the dietary study to the location(s) from which the moulds. Casts were made from these moulds using EpoTek 320 LV Black epoxy museum specimens we analysed were collected (specimen provenance data, where resin mixed to the manufacturer’s instructions. Resin was cured for 24 h under 200 known, is included in Supplementary Data 1). Taxa that had not been subjected to kPa (2 Bar/30 psi) of pressure (Protima Pressure Tank 10L) to improve casting ecological studies that provided volumetric or frequency breakdowns of diet were 32,41,45,48 quality. Small casts were mounted onto 12.7 mm SEM stubs using President Jet not sampled for study. Many reptiles exhibit ontogenetic shifts in diet , polyvinylsiloxane with the labial, non-occluding surfaces orientated dorsally to thus similarly sized specimens from extant species were sampled where possible to optimise data acquisition. All casts were sputter coated with gold for three minutes minimise effects of this confounding variable. Lower jaw lengths, measured in (SC650, Bio-Rad, Hercules, CA, USA) to optimise capture of surface texture data. lateral view, were used as a proxy for overall body size. For extant reptiles, lower Replicas produced using these methods are statistically indistinguishable from jaw lengths were measured from the anterior tip of the dentary to the posterior original tooth surfaces65. margin of the surangular. Bat jaw lengths were measured from the anterior to posterior tip of the dentary. Specimen availability enabled Crocodylus porosus to be sampled as two separate life-history groups for analysis. Specimens with lower jaw Surface texture data acquisition. Surface texture data acquisition follows stan- lengths of less than 50 cm were classified as juveniles; specimens with jaw lengths dard laboratory protocols16–18,65,66. Data were captured using an Alicona Infinite exceeding 50 cm were classified as adults18,61. Focus microscope G4b (IFM; Alicona GmbH, Graz, Austria; software version 5.1), Extant reptiles and bats were assigned to dietary guilds based on the relative using a ×100 objective lens, producing a field of view of 146 × 100 μm. Lateral and ‘intractability’ (roughly equivalent to hardness) of prey as food18. Guilds include vertical resolution were set at 440 and 20 nm, respectively. Casts were orientated so carnivores (tetrapod consumers); piscivores (fish consumers); ‘harder’ invertebrate labial surfaces were perpendicular to the axis of the objective lens. consumers (invertebrates with hard exoskeletons, e.g. beetles, crustaceans and All 3D data files were processed using Alicona IFM software (version 5.1) to shelled gastropods); ‘softer’ invertebrate consumers (invertebrates with less hard remove dirt particles from tooth surfaces and anomalous data points (spikes) by exoskeletons, e.g. crickets, grasshoppers, dragonflies, damselflies and ants); ‘softest’ manual deletion. Data were levelled (subtraction of least-squares plane) to remove invertebrate consumers (invertebrates with soft exoskeletons, e.g. invertebrate variation caused by differences in tooth surface orientation at the time of data larvae, butterflies, moths, arachnids and millipedes); omnivores (combination of capture. Files were exported as .sur files and imported into Surfstand (software plant and matter). These guilds do not represent the entire dietary diversity version 5.0.0 Centre for Precision Technologies, University of Huddersfield, West of the extant , but were selected because of their ecological relevance to Yorkshire, UK). Scale-limited surfaces were generated through application of a pterosaur diet. For example, many New World bat species are nectarivores, fifth-order robust polynomial to remove gross tooth form and a robust Gaussian 62 fi λ = 17,67 19 sanguivores (blood feeders) or obligate frugivores . These diets have never been lter (wavelength c 0.025 mm) . ISO 25178-2 areal texture parameters

6 NATURE COMMUNICATIONS | (2020) 11:5293 | https://doi.org/10.1038/s41467-020-19022-2 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19022-2 ARTICLE were then generated from each scale-limited surface. Descriptions of ISO Code availability parameters can be found in Supplementary Data 4 (from ref. 18). The code used for the dietary evolution reconstructions can be found in the R file Supplementary Code 1, and is available in Zenodo (https://doi.org/10.5281/ 79 DMTA statistical analyses. Log-transformed texture data were used for analyses as zenodo.4018876) . some of the texture parameters were non-normally distributed (Shapiro-Wilk, P > 0.05). The parameter Ssk was excluded from analysis as it contains negative values Received: 8 January 2020; Accepted: 25 September 2020; and thus cannot be log-transformed. To test the hypotheses that microwear differs between reptiles and bats from different dietary guilds, analysis of variance (ANOVA) with pairwise testing (Tukey HSD) was used for each texture parameter. Where homogeneity of variance tests revealed evidence of unequal variances, Welch analysis of variance was used. Principal component analysis (PCA) was used to explore texture parameters exhibiting significant differences between reptile and bat dietary guilds. ANOVA with pairwise testing was used on the PC axis 1 and 2 values of dietary References guilds to determine whether guilds occupy different areas of multivariate space along 1. Witton, M. P. Pterosaurs: Natural History, Evolution, Anatomy (Princeton these axes (see ref. 18 for results for reptiles). To test the hypotheses that reptile and Univ. Press, 2013). bat microwear differences are determined by dietary differences, we used Spearman’s 2. Barrett, P. M., Butler, R. J., Edwards, N. P. & Milner, A. R. Pterosaur rank to test for correlations between PC axes 1 and 2 and dietary characteristics. distribution in time and space: an atlas. Zitteliana 28,61–107 (2008). Additional analyses were employed to independently test for subtle microtextural 3. Dean, C. D., Mannion, P. D. & Butler, R. J. Preservational bias controls the differences between dietary guilds. Texture parameters were ranked by guild based on fossil record of pterosaurs. Palaeontology 59, 225–247 (2016). the average value for the guild; matched pairs t-tests were used to compare the profiles 4. Bestwick, J., Unwin, D. M., Butler, R. J., Henderson, D. M. & Purnell, M. A. of average parameter values between guilds (see ref. 18 for results for reptiles, and Pterosaur dietary hypotheses: a review of ideas and approaches. Biol. Rev. 93, Supplementary Tables 1 and 2 for results for bats). 2021–2048 (2018). A Benjamini–Hochberg (B–H) procedure was used to account for the 5. Wang, X. & Zhou, Z. Pterosaur assemblages of the Jehol Biota and their possibility of inflated I error rates associated with multiple comparisons68. implication for the Early Cretaceous pterosaur radiation. Geol. J. 41, 405–418 The false discovery rate was set at 0.05. The B–H procedure was not needed for the (2006). Tukey HSD tests as it already accounts for inflated Type I error rates69. 6. Bennett, S. C. A statistical study of Rhamphorhynchus from the Solnhofen Pterosaur microwear data were independently projected onto the axes of the limestone of Germany: -classes of a single large species. J. Paleontol. 69, reptile and bat analyses. Correlation of Rhamphorhynchus lower jaw lengths with 569–580 (1995). ’ PC 1 values were tested using Spearman s rank tests. All DMTA analyses were 7. Prondvai, E., Stein, K., Ősi, A. & Sander, M. P. Life history of – performed with JMP Pro 12 (SAS Institute, Cary, NC, USA) except for the B H Rhamphorhynchus inferred from bone and the diversity of 70 procedure , which used Microsoft Excel (www.biostathandbook.com/ pterosaurian growth strategies. PLoS ONE 7, e31392 (2012). multiplecomparisons.html). 8. Bennett, S. C. New smallest specimen of the pterosaur and ontogenetic niches in pterosaurs. J. Vert. Paleontol. 92, 254–271 (2017). Pterosaur dietary evolution. Pterosauria is defined as the most recent common 9. Benson, R. B., Frigot, R. A., Goswami, A., Andres, B. & Butler, R. J. ancestor of buffarinii and northropi and all its des- Competition and constraint drove Cope’s rule in the evolution of giant flying cendants4. Several competing phylogenies exist for Pterosauria which show similar, reptiles. Nat. Comm. 5, 3567 (2014). but disputed, taxonomic contents of clades and grades of morphologically similar 10. Nesbitt, S. J. The early evolution of archosaurs: relationships and the origin of taxa4. We therefore used three independent phylogenies, Lü et al.33, Andres and major clades. Bull. Am. Mus. Nat. Hist. 352,1–292 (2011). Myers34 and Wang et al.35, to visualise reconstructed pterosaur diets from DMTA 11. Witton, M. P. in New Perspectives on Pterosaur Palaeobiology (eds. Hone, D. in evolutionary contexts. These phylogenies have extensive taxonomic sampling W. E., Witton, M. P. & Martill, D. M.) Vol. 455, 7–23 (Geol. Soc. Lond. Spec. across Pterosauria and contain the majority of taxa in this study. See Supple- Pub., 2018). 34 35 mentary Note 3 for Andres and Myers and Wang et al. results. A time- 12. Ősi, A. Feeding-related characters in basal pterosaurs: implications for jaw calibrated tree was constructed from each phylogeny using the DatePhylo function mechanism, dental function and diet. Lethaia 44, 136–152 (2011). 71 72 of the R package strap, version 1.4 (ref. ) (R version 3.2.3 (ref. )). First and last 13. Zhou, C.-F., Gao, K.-Q., Yi, H., Xue, J., Li, Q. & Fox, R. C. Earliest filter- 3 appearance dates were obtained from Dean et al. and the Paleobiology Database feeding pterosaur from the Jurassic of China and ecological evolution of ‘ ’ (PBDB: www.paleobiodb.org). Branches were scaled using the equal method . R. Soc. Open Sci. 4, 160672 (2017). which eliminates zero-length branches by sharing duration from the more basal fi 71 14. Ungar, P. S., Brown, C. A., Bergstrom, T. S. & Walker, A. Quanti cation of non-zero-length branches . Time-scaled trees were pruned in R to include only dental microwear by tandem scanning confocal microscopy and scale- pterosaurs subjected to DMTA. Pruning has no effect on the topological rela- sensitive fractal analysis. Scanning 25, 185–193 (2003). tionships of remaining pterosaurs. Missing pterosaurs (i.e. species included in the 15. Daegling, D. J. et al. Viewpoints: feeding mechanics, diet and dietary DMTA dataset but not in the phylogenetic hypothesis) were included by hand adaptations in early hominins. Am. J. Phys. Anthropol. 151, 356–371 (2013). prior to time-scaling based on information from the literature on their likely 16. Purnell, M., Seehausen, O. & Galis, F. Quantitative three-dimensional phylogenetic position (e.g. assignments to particular clades in published systematic microtextural analyses of tooth wear as a tool for dietary discrimination in assessments). To minimise confounding effects of ontogeny, only the largest spe- fi fishes. J. R. Soc. Interface 9, 2225–2233 (2012). cimen of each pterosaur taxa was included in the analysis, identi ed by lower jaw fi length in lateral view from the anterior tip of the dentary to the posterior margin of 17. Purnell, M. A. & Darras, L. P. G. 3D tooth microwear texture analysis in shes the surangular. Each tree was independently projected onto the first two PC axes of as a test of dietary hypotheses of durophagy. Surf. Topogr. Metrol. Prop. 4, the reptile texture-dietary space to generate pterosaur phylo-texture-dietary spaces 014006 (2016). using the phylomorphospace function of the R package phytools, version 0.6-60 18. Bestwick, J., Unwin, D. M. & Purnell, M. A. Dietary differences in archosaur (ref. 73). Phylogenetic signals in pterosaur microwear along PCs 1 and 2 were tested and lepidosaur reptiles revealed by dental microwear textural analysis. Sci. using the phylosig function in phytools73. Pterosaur PC 1 and PC 2 values were Rep. 9, 11691 (2019). compared to time from the root of the time-calibrated trees in millions of years to 19. ISO 25178-2. Geometrical Product Specifications (GPS)—Surface Texture: test for shifts through phylo-texture-dietary space over evolutionary time. Ances- Areal—Part 2: Terms, Definitions and Surface Texture Parameters tral character-state reconstructions of pterosaur diets were performed by inde- (International Organisation for Standardisation) 1–42 (2012). pendently mapping pterosaur PC 1 and PC 2 values as continuous characters onto 20. Winkler, D. E., Schulz-Kornas, E., Kaiser, T. M. & Tütken, T. Dental each pruned tree using the contMap function in phytools73. Ancestral PC 1 and PC microwear texture reflects dietary tendencies in extant Lepidosauria despite 2 values (and the variance and 95% confidence intervals) were calculated for each their limited use of oral food processing. Proc. R. Soc. B. 286, 20190544 (2019). node within the three phylogenies. Pterosaur skull diagrams in Figs. 1 and 3–4 and 21. Frey, E. & Tischlinger, H. The pterosaur Rhamphorhynchus,a Supplementary Figs. 4 and 6–8 were traced from ref. 4 under a Creative Commons frequent victim of the ganoid fish ? PLoS ONE 7, e31945 Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/) (2012). except Anhanguera, Austriadactylus, Boreopterus, Coloborhynchus and Jiang- 22. Wellnhofer, P. Die (Pterosauria) der Oberjura- – changnathus, which were redrawn from refs. 74 78, respectively. Plattenkalke Süddeutschlands, Teil III: Palökologie und Stammesgeschichte. Palaeontogr. Abt. A 1–3,1–30 (1975). Reporting summary. Further information on research design is available in the Nature 23. Hone, D., Henderson, D. M., Therrien, F. & Habib, M. B. A specimen of Research Reporting Summary linked to this article. Rhamphorhynchus with soft tissue preservation, stomach contents and a putative coprolite. PeerJ 3, e1191 (2015). 24. Wellnhofer, P. The Illustrated Encyclopaedia of Pterosaurs (Salamander Books, Data availability 1991). All the data analysed in this study are publically available in Zenodo (https://doi.org/ 25. Witton, M. P. New insights into the skull of Istiodactylus latidens 10.5281/zenodo.4018876)79 and within the Supplementary Information files. (Ornithocheiroidea, Pterodactyloidea). PLoS ONE 7, e33170 (2012).

NATURE COMMUNICATIONS | (2020) 11:5293 | https://doi.org/10.1038/s41467-020-19022-2 | www.nature.com/naturecommunications 7 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19022-2

26. Dalla Vecchia, F. M. in Anatomy, Phylogeny and Palaeobiology of Early 59. Humphrey, S. R., Bonaccorso, F. J. & Zinn, T. J. Guild structure of surface- Archosaurs and their Kin (eds. Nesbitt, S.J., Desojo, J.B. & Irmis, R.B.) 379, gleaning bats in Panamá. Ecology 64, 284–294 (1983). 119–155 (Geol. Soc. Lond. Spec. Pub., 2013). 60. Bonato, V., Facure, K. G. & Uidea, W. Food habits of bats of subfamily 27. Veldmeijer, A. J., Signore, M. & Bucci, E. in Predation in Organisms—A Vampyrinae in Brazil. J. Mammal. 85, 708–713 (2004). Distinct Phenomenon (ed. Elewa, A.M.T.) 295–308 (Springer, 2007). 61. Fukuda, Y., Saafeld, K., Lindner, G. & Nichols, T. Estimation of total length 28. Unwin, D. M. & Martill, D. M. in The Crato Fossil Beds of Brazil: Window into from head length of saltwater crocodiles (Crocodylus porosus) in the Northern an Ancient World (eds. Martill, D. M., Bechly, G. & Loveridge, B.) 475–524. Territory, Australia. J. Herpetol. 47,34–40 (2013). (Cambridge Univ. Press, 2007). 62. Santana, S. E., Dumont, E. R. & Davis, J. L. Mechanics of bite force production 29. Henderson, D. M. in New Perspectives on Pterosaur Palaeobiology (eds. Hone, D. W. and its relationship to diet in bats. Funct. Ecol. 24, 776–784 (2010). E., Witton, M. P. & Martill, D. M.) 455, 25–44 (Geol. Soc. Lond. Spec. Pub., 2018). 63. Williams, V. S. & Doyle, A. M. Cleaning fossil tooth surfaces for microwear 30. Unwin, D. M. The Pterosaurs from Deep Time (Pi Press, 2006). analysis: use of solvent gels to remove resistant consolidant. Palaeontol. 31. Hoffmann, R., Bestwick, J., Berndt, G., Fuchs, D. & Klug, C. Pterosaurs ate Electron. 13, 2T (2010). soft-bodies (Coleoidea). Sci. Rep. 10, 1230 (2020). 64. Schubert, B. W. & Ungar, P. S. Wear facets and enamel spalling in 32. Platt, S. G., Thorbjarnarson, J. B., Rainwater, T. R. & Martin, D. R. Diet of the tyrannosaurid . Acta Palaeontol. Pol. 50,93–99 (2005). American Crocodile (Crocodylus acutus) in marine environments of Coastal 65. Goodall, R. H., Darras, L. P. & Purnell, M. A. Accuracy and precision of Belize. J. Herpet 47,1–10 (2013). silicon based impression media for quantitative areal texture analysis. Sci. Rep. 33. Lü, J., Kundrát, M. & Shen, C. New material of the pterosaur 5, 10800 (2015). Gladocephaloideus Lü et al., 2012 from the Early Cretaceous of Liaoning 66. Purnell, M. A., Crumpton, N., Gill, P. G., Jones, G. & Rayfield, E. J. Within- Province, China, with comments on its systematic position. PLoS ONE 11, guild dietary discrimination from 3-D textural analysis of tooth microwear in e0154888 (2016). insectivorous mammals. J. Zool. 291, 249–257 (2013). 34. Andres, B. & Myers, T. S. Lone star pterosaurs. Earth Env. Sci. Trans. R. Soc. 67. Schulz, E., Calandra, I. & Kaiser, T. M. Feeding ecology and chewing Edinburgh. 103, 383–398 (2013). mechanics in hoofed mammals: 3D tribology of enamel wear. Wear 300, 35. Wang, X. et al. New evidence from China for the nature of the pterosaur 169–179 (2013). evolutionary transition. Sci. Rep. 7, 42763 (2017). 68. Benjamini, Y. & Hochberg, Y. Controlling the false discovery rate: a practical 36. Andres, B. Death and taxa: the shape of pterosaur evolution. In The 5th and powerful approach to multiple testing. J. R. Stat. Soc. B. 57, 289–300 International Symposium on Pterosaurs (Univ. Portsmouth, UK, 2015). (1995). 37. Wang, M., Wang, X., Wang, Y. & Zhou, Z. A new basal bird from China with 69. Purnell, M. A., Goodall, R. H., Thomson, S. & Matthews, C. J. D. Tooth implications for morphological diversity in early birds. Sci. Rep. 6, 19700 (2016). microwear texture in odontocete whales: variation with tooth characteristics 38. Losos, J. B. & Greene, H. W. Ecological and evolutionary implications of diet and implications for dietary analysis. Biosurf. Biotribol. 3, 184–195 (2017). in monitor lizards. Biol. J. Linn. Soc. 35, 379–409 (1988). 70. McDonald, J. H. Handbook of Biological Sciences 3rd edn. (Sparky House 39. Nifong, J. C. et al. Animal-borne imaging reveals novel insights into the Publishing, 2014). foraging behaviors and diel activity of a large-bodied apex predator, the 71. Bell, M. A. & Lloyd, G. T. strap: an R package for plotting phylogenies against American alligator (Alligator mississippiensis). PLoS ONE 9, e83953 (2014). stratigraphy and assessing their stratigraphic congruence. Palaeontology 58, 40. Chatterjee, S. & Templin, R. J. Posture, Locomotion, and Paleoecology of 379–389 (2014). Pterosaurs. Vol. 376, 1–64 (Geol. Soc. Am. Spec. Pub, 2004).. 72. R Development Core Team. R: A Language and Environment for Statistical 41. Delany, M. F., Linda, S. B. & Moore, C. T. Diet and condition of American Computing (R Foundation for Statistical Computing, Vienna, Austria, 2010). in 4 Florida lakes. Proc. Annual Conference on Southeastern 73. Revell, L. J. Phytools: an R package for phylogenetic comparative biology (and Association of and Wildlife Agencies (eds Eversole, R. G., Wong, K. C., other things). Methods Ecol. Evol. 3, 217–223 (2012). Chamberlain, M. & Rohwer, F.) Vol. 53, 375–389 (1999). 74. Bantim, R. A. M., Saraiva, A. A. F. & Sayão, J. M. Skull variation and the shape 42. Laverty, T. M. & Dobson, A. P. Dietary overlap between black caimans and of the sagittal premaxillary crest in anhanguerid pterosaurs (Pterosauria, spectacled caimans in the Peruvian Amazon. Herpetologica 69,91–101 (2013). Pterodactyloidea) from the Araripe Basin, Northeast Brazil. Hist. Biol. 27, 43. Thorbjarnarson, J. B. The status and ecology of the American crocodile in 656–664 (2014). Haiti. Bull. Fla. Mus. Biol. Sci. 33,1–86 (1988). 75. Dalla Vecchia, F. M. The first Italian specimen of Austriadactylus cristatus 44. Wallace, K. M. & Leslie, A. J. Diet of the Nile Crocodile (Crocodylus niloticus) (Diapsida, Pterosauria) from the Norian (Upper Triassic) of the carnic in the Okavango Delta. Botsw. J. Herpet. 42, 361–368 (2008). Prealps. Riv. Ital. Paleontol. S. 115, 291–304 (2009). 45. Sah, S. A. M. & Stuebing, R. B. Diet, growth and movements of juvenile 76. Jiang, S.-X., Wang, X.-L., Meng, X. & Cheng, X. A new boreopterid pterosaur crocodiles Crocodylus porsus Schneider in the Klias river, Sabah, Malaysia. J. from the Lower Cretaceous of western Liaoning, China, with a reassessment of Trop. Ecol. 12, 651–662 (1996). the phylogenetic relationships of the Boreopteridae. J. Paleontol. 88, 823–828 46. Taylor, J. A. The foods and feeding habits of subadult Crocodylus porosus (2014). Schneider in Northern Australia. Aust. Wildl. Res. 6, 347–359 (1979). 77. Martill, D. M. First occurrence of the pterosaur Coloborhynchus (Pterosauria, 47. Hussain, S. A. Ecology of Gharial (Gavialis gangeticus) in National Chambal ) from the Wessex Formation (Lower Cretaceous) of the Isle Sanctuary, India. MPhil thesis, Aligarh Muslim University (1991). of Wight, . Proc. Geol. Assoc. 126, 377–380 (2015). 48. Auffenberg, W. The Behavioral Ecology of the Komodo Monitor (Univ. Press, 78. Cheng, X., Wang, X., Jiang, S. & Kellner, A. W. A. A new scaphognathid Florida, 1981). pterosaur from western Liaoning. China Hist. Biol. 24, 101–111 (2012). 49. Dalhuijsen, K., Branch, W. R. & Alexander, G. J. A comparative analysis of the 79. Bestwick, J., Unwin, D. M., Butler, R. J. & Purnell, M. A. Dietary diversity and diets of Varanus albigularis and Varanus niloticus in South Africa. Afr. Zool. evolution of the earliest flying vertebrates revealed by dental microwear 49,83–93 (2014). texture analysis. Data Sets. Zenodo https://doi.org/10.5281/zenodo.4018876 50. Auffenberg, W. Gray’s Monitor Lizard (Univ. Press, Florida, 1988). (2020). 51. Greene, H. W. Diet and arboreality in the emerald monitor, Varanus prasinus, with comments on the study of adaptation. Fieldiana Zool. 31,1–12 (1986). 52. Rahman, K. M. M., Rakhimov, I. I. & Khan, M. M. H. Activity budgets and Acknowledgements dietary investigations of Varanus salvator (Reptilia: Varanidae) in Karamjal Thanks to P. Campbell, M. Carnall, T. Davidson, E. Frey, S. Harris, J. Hay, S.-X. Jiang, D. ecotourism spot of Bangladesh Sundarbans mangrove forest. Basic Appl. Lunde, R. Miguez, O. Rauhut, A. Resetar, R. Schoch, T. Schossleitner, A. Schulp, C. Herpet 31,45–56 (2017). Sheehy, L. Steel, B. Strilisky, A. Wynn and C.-F. Zhou for specimen access. Thanks to S. 53. Almenar, D., Aihartza, J., Goiti, U., Salsamendi, E. & Garin, I. Diet and prey Gabbott and D. Henderson for comments. This work was funded by an NERC stu- selection in the trawling long-fingered bat. J. Zool. 274, 340–348 (2008). dentship awarded through the Central England NERC Training Alliance (CENTA; grant 54. Berge, L. Resource Partitioning between the Cryptic Species Brandt’s bat reference NE/L002493/1) and the University of Leicester, and by a Palaeontological (Myotis brandtii) and the Whiskered Bat (M. mystacinus) in the UK. PhD Association Sylvester-Bradley Award (PA-SB201701) to J.B. J.B. was also supported by a thesis, Univ. Bristol (2007). Leverhulme Trust Research Project Grant (RPG-2019-364) during the completion of the 55. Brooke, A. P. Diet of the fishing bat, Noctilio leporinus (Chiroptera: project. Noctilionidae). J. Mammal. 75, 212–218 (1994). 56. Jones, G. Flight peformance, echolocation and foraging behaviour in noctule Author contributions bats Nyctalus noctula. J. Zool. Soc. Lond. 237, 303–312 (1995). M.A.P. and D.M.U. conceived the study. All authors contributed to the analytical design. 57. Rydell, J. Food habits of northern (Eptesicus nilssoni) and brown long-eared J.B. collected the data. J.B. analysed the data with input from M.A.P. and R.J.B. J.B. wrote (Plecotus auritus) bats in Sweden. Holarct. Ecol. 12,16–20 (1989). the paper with contributions from all authors. 58. Flanders, F. & Jones, G. Roost use, ranging behavior, and diet of greater horseshoe bats (Rhinolophus ferrumequinum) using a transitional roost. J. Mammal. 90, 888–896 (2009). Competing interests The authors declare no competing interests.

8 NATURE COMMUNICATIONS | (2020) 11:5293 | https://doi.org/10.1038/s41467-020-19022-2 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19022-2 ARTICLE

Additional information Open Access This article is licensed under a Creative Commons Supplementary information is available for this paper at https://doi.org/10.1038/s41467- Attribution 4.0 International License, which permits use, sharing, 020-19022-2. adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Correspondence and requests for materials should be addressed to J.B. or M.A.P. Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless Peer review information Nature Communications thanks Ivan Calandra, Danielle Fraser indicated otherwise in a credit line to the material. If material is not included in the and Eberhard Frey for their contribution to the peer review of this work. Peer reviewer article’s Creative Commons license and your intended use is not permitted by statutory reports are available. regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ Reprints and permission information is available at http://www.nature.com/reprints licenses/by/4.0/.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. © The Author(s) 2020

NATURE COMMUNICATIONS | (2020) 11:5293 | https://doi.org/10.1038/s41467-020-19022-2 | www.nature.com/naturecommunications 9